Feasibility study of a vision-based landing system for unmanned fixed-wing aircraft
McCarthy, Tyler B.
Yakimenko, Oleg A.
Papoulias, Fotis A.
MetadataShow full item record
Successful landing of an autonomous unmanned aerial vehicle requires a high degree of accuracy and efficient, real-time processing. This research applies systems engineering concepts to investigate the feasibility of applying computer vision techniques and visual feedback in the control loop for an autonomous system. This thesis examines the framework and performance of an algorithm designed to detect and track a runway in images captured from a camera onboard an aircraft during the final approach and landing stages of flight. Using a series of image processing techniques to localize the runway and the Hough transformation for line detection, the algorithm is capable of detecting the edges of a runway with over 96 percent accuracy through 3000 test images. The operating conditions for this algorithm include any scenario in which visual flight rules apply. Additionally, the system will perform with runways that comply with Federal Aviation Administration regulations. Future applications of this algorithm should include aircraft attitude and pose estimation as well as full integration into an autonomous aircraft control system.
Approved for public release; distribution is unlimited
Showing items related by title, author, creator and subject.
McCarthy, Tyler (2017-04-12);This presentation outlines an approach to develop a robust, real-time runway detection and tracking system for an unmanned aerial vehicle (UAV) using computer vision and standard guidance, navigation, and control (GNC) ...
Bollino, Kevin P. (Monterey, California.: Naval Postgraduate School, 2006., 2006-12);Creating simplicity out of complexity, this research abandons the traditional guidance and control architecture for aerospace vehicles and embraces a revolutionary concept based on the principles of nonlinear optimal ...
Morsy, Khaled Ahmed (Monterey, California. Naval Postgraduate School, 1997-09);The problem discussed in this dissertation is the development of an efficient method for visual navigation of autonomous vehicles. The approach is to significantly reduce the expensive computational time of landmark detection ...